If this problem is as big as it sounds, what are the repercussions?
If I flip a coin one million times and only publish the ~500,000 results that show coin toss coming up as "heads," I could lead people to believe that a fair coin will always land on heads.
That's how misleading this bias can be.
But anyone can toss a fair coin, so I'd be found out pretty quick. But with more complicated and murky stuff like psychology this can be a huge problem, especially if people begin using this shaky work as foundation for things like: public policy, how to live their lives, how to do work, etc.
After all they can't just toss a coin to check my work on the more complicated stuff.
To take your fair coin example, publication bias is when the only studies that get published are the ones that show an unfairness in the coin – the rest of the studies that simply find 50% heads and 50% tails (aka a fair coin) don't get published because they're not interesting.
I don't think there's a way to distinguish between studies that have lost funding because there are genuinely less funds to go around, and foo experiment just happened to be one of the ones that didn't make the cut, and studies that have had the funding pulled for harmful reasons (Like negative results).
Mainly because it's devilishly easy to mask; you'd just reduce funding to research and redistribute it into something plausible like marketing. Of course this isn't feasible for lots of studies producing negative results.
Of course, one way to counter that is to insist that the data is published regardless, but I'm sure that the data could be hidden with clever use of NDAs and court arguments along the lines of "We need to prohibit the distribution of company assets".
But something else came to my mind now:
> Of course, one way to counter that is to insist that the data is published regardless, but I'm sure that the data could be hidden with clever use of NDAs and court arguments along the lines of "We need to prohibit the distribution of company assets".
I find myself more and more treating complexity as a proxy for dishonesty. Companies that are trustworthy seem to have a relatively simple business model that they don't try to hide. The more convoluted it is, the more likely it is someone is scamming you. Wonder if that could be used as an effective metric - the more complex the reason someone weasels away from publishing the data, the more their "reputability" score goes down?
Fraud requires intent. Now, this extreme of bias would typically require said intent, sure. But let's say the experimenter has anterograde amnesia.
They get an idea, flip a coin - oh hey, it's heads! They publish the interesting result. They get an idea, flip a coin - oh, it was tails, nevermind that theory. They get an idea... each time, they don't remember that they've already attempted the theory. No intent, no fraud, just bias.
The good news is this is a pretty unlikely extreme in an individual. Anterograde amnesia is rare. The bad news is nobody can remember that someone else attempted a theory and came up with a negative result that they didn't bother to publish, and even individuals will have a hard time perfectly accounting for and preventing their own bias within a single study. The even worse news is that fraud exists as well.
There was an econtalk episode late last year with Brian Nosek about the Reproducability Project (http://www.econtalk.org/archives/2015/11/brian_nosek_on.html). They went the further step of analyzing the wider field to see how bad the biases are. Previous attempts to call out replication (priming theory) did not go well, leading to a lot of hostility. Nosek's process is a bit more friendly, by inviting the original authors's critique early, before data collection. I figure it's a bit harder to blast the auditors for considering a replication study without appearing as though you doubt the effect is real yourself. They still got a bit of pushback, but nowhere near as acrimonious.
But psi researchers these days jump through far more hoops, run far more replications, spend far more time looking for weaknesses and alternative explanations, garner far more scrutiny and second guessing. To call that a 'fake field of research' when it is more rigorous is disingenuous.
Negative results do get published regularly. And more replications take place. It is quite hilarious really, because it makes people panic about what is wrong with the methods we apply across the board in science. Skeptics just raise the bar abritrarily and shift the goalposts as a means of sweeping things under the rug while calling it scientific rigour.
We have also now gotten to the point that the fact that we commonly have failed replications (across all social science), it means the whole premise of a line of research must be completely flawed. But we should expect failed replications - for one, it's impossible to repeat a set of circumstances exactly. Many replications aren't true replications.
So what's a good study you can point me at? The nice thing about psi testing is that you can get a result sample every few seconds. It should be easy to turn even a moderately small effect into a five sigma result that's immune to publication bias.
"During 2013 and 2014, a total of 1479 people from 77 countries contributed 2985 test sessions. Over the same period, 5738 sessions were run as controls by a computer programmed to simulate human participants. The results showed that with human observers the fringe visibility at the center of the interference pattern deviated from a null effect by 5.72 sigma (p = 1.05 × 10−8), with the direction of the deviation conforming to the observers' intentions. The same analysis applied to the control data resulted in an overall deviation of −0.17 sigma."
There are some presentations on Youtube that go into the details of the double-slit line of experiments.
But that's the point of Scott's article. We have more than enough reasons to believe the whole field is bullshit (did you see a working practical application of that research?), but because it does all the moves and rituals of Real Science, it forms a perfect baseline for how many positive results we can expect with pure cargo-cult science.
The explanation that psi-researchers are just good at finding the appearance of signal from noise, small as it may be, is fine, but then personally, I'd like to see the same vigour for explaining what the signal actually is. And people that have a firm belief that any 'psi'-explanation can't possibly true (and this a pretty arbitrary notion at that, as psi would just follow conventional physics and adhere to normal biological mechanisms) just aren't going to accept any other explanation. So they look for issues with the statistics used.
The notition that the whole field is BS, as you put it, requires all the researchers in involved to be utterly mad, stupid or deceitful. That is a claim easily made by skeptics, a thing that they have no trouble agreeing to. That view is not born out of scientific honesty, it's just another form of cargo-cult thinking.
In this way, paranomral reserach is a good "control group" for science because it uses methods similar to those of psychology and the social sciences.
In medicine all kinds of things were done to avoid publication bias (most notably trials registers). It still is a big problem, because the countermeasures aren't enforced. In most other areas people don't even try to counter it, which means it's probably much worse and we don't know how bad it is. It's a major reason why so much of science is most likely false.
I suggest you have a look at the work of Ben Goldacre, he's been a major voice in highlighting this issue. He's done a ted talk introducing the problem and later has written a book (Bad Pharma). He's only focussing on medical trials, but you always have to think: "This is happening in all areas of science".
So in support and just for reference, some articles:
http://journals.plos.org/plosmedicine/article?id=10.1371/jou... (Why Most Published Research Findings Are False)
http://www.scientificamerican.com/article/an-epidemic-of-fal...
http://www.economist.com/news/briefing/21588057-scientists-t...
http://www.economist.com/news/leaders/21588069-scientific-re...
http://www.economist.com/blogs/babbage/2013/10/science-wrong
In particular, publication bias isn't quite as bad as the garden-of-forking-paths issue.